The awful toll of poverty across the world should make the strongest possible claim on the conscience of the rich West, citizens and governments alike. But the numbers are so huge, they are paralyzing. Over a billion people, more than three times the population of the United States, live in extreme poverty. What is "extreme poverty"? It is the kind of poverty that no American, however poor, ever experiences. It means that you barely have the means to stay alive, even when times are good. If times are not good and you become ill, or have to cope with a drought or a bad harvest, you are quite likely to die. In sub-Saharan Africa today, an infant's chance of surviving to the age of 65 is roughly one in three.

If wealthy countries could do something to attack this scourge, the moral case for action would be irresistible. But can they? The record of foreign aid in advancing economic development and reducing poverty in the poorest countries has been weak. Since the 1960s, billions of dollars have been spent to no purpose. As a result, in many Western capitals, aid fatigue set in long ago.

Rich-country donors still talk a good fight and go through the motions, but with no great conviction—and in most cases, their spending on aid is miserly. This week, the United Nations published a report, "Investing in Development: A Practical Plan to Achieve the Millennium Development Goals," which is intended, among other things, to rekindle enthusiasm for development aid and to persuade the rich countries to spend far more on it.

This document might actually have some effect. If it does, this will partly be because the U.N. chose Jeffrey Sachs to lead the team that produced it. Sachs, a professor at Columbia University, is a distinguished economist, an adviser to numerous poor countries, and a globally recognized authority on development. He is also a man of limitless energy and awesome self-belief. It is not often that the U.N. publishes a document that so stirs the passions, and that puts such moral pressure on whoever reads it to demand action of their governments.

In doing both, this new report may change the terms of the debate about development. As a piece of advocacy, therefore, it could well succeed. But is its analysis true—or is the report, assuming that it does succeed in spurring action, the prelude to another surge of meaningless aid yielding no results?

Sachs and his team squarely confront the idea that aid has failed in the past. This conventional wisdom, they argue, is a myth. First, the team says, studies that naively try to correlate aid with economic growth are misleading. A lot of aid is sent to countries that are in distress following natural disasters—such as the Indian Ocean tsunami—or in response to other humanitarian emergencies. Countries facing such crises will typically suffer setbacks to economic growth at the same time. For this reason, aid will be speciously correlated with low growth; this does not mean that aid does not work. When you separate humanitarian relief from the rest, and look to see whether development aid promotes growth, the results are better.

Also, Sachs and his team argue, a great deal of past aid has been badly designed. Different donors have pursued different aid strategies, and have failed to coordinate with one another. Aid commitments have been short-term, making it difficult to meet the recurrent costs of longer-term development projects. Specific development needs vary a lot from country to country, the report argues, something that donors have tended to ignore. Design the aid well, say Sachs and his team, and the results will be good—provided that one other crucial condition is met.

Aid works well only in countries that are reasonably well governed. This is one of the clearest findings from recent research on development. Corruption, official incompetence, and the failure to protect basic property rights are capable of rendering aid not just useless but harmful, as when it enriches corrupt leaderships and strengthens their hold on power. This poses quite a problem for advocates of aid, especially of aid for the poorest countries, because those countries are typically badly governed. It is a dilemma: The countries that need aid most are often the ones most likely to waste it; the ones that can best use it need it less, if at all.

The report agrees that reasonably good government is vital for certain kinds of aid to work, but it says that some of the poorest countries do in fact meet the governance test. These countries—which in Africa would include Ghana, Mozambique, Tanzania, and Uganda, and which might number at least a dozen worldwide—should be fast-tracked for much more aid, it says. And the report is full of suggestions about how the money should be spent. The results, it says, would be impressive. Once aid was seen to work, support for it in the West would increase. Governments might even be willing to keep the promise they have been making for 35 years to give 0.7 percent of their national income each year in aid. (Most fall far short; America gives around 0.2 percent.) And the "tough-love" approach of being far more generous to well-run countries would encourage competition among other poor countries to improve their standards of governance.

It all sounds fine, but does it really add up? Some of it does, but there are problems. The emphasis on good government as a precondition is welcome and, so far as the U.N. is concerned, something of a departure. But the report itself flinches somewhat from the implications of this point—and, in practice, the U.N. would do so even more violently. The report blurs the line between very poor countries capable of using aid well and very poor countries capable only of wasting it. For instance, it apologizes for bad governance in most of sub-Saharan Africa by saying that many African countries are in fact pretty well governed, allowing for their level of income. So what? The thing that matters is not whether countries are well run in any relative sense, but whether they are sufficiently well run to use aid well. This notion of "relatively well governed" is an evasion.

In the same vein, the report also talks of countries that are "potentially well governed"—countries, it says, with governments that have the will to rule well but not the means; countries that would be well governed given a bit of help from outside. Again, this seems a very dubious concept. Certainly, using aid to improve governance would be a fine idea if it worked. But this has been tried many times with notably little success. Meanwhile, "potentially well governed" just is not good enough.

Even letting that go, the report's list of "potentially well-governed" countries is perplexing. It includes five of the seven most corrupt countries in the world, as judged by Transparency International (a nongovernmental organization that ranks countries for this purpose): Azerbaijan, Bangladesh, Chad, Nigeria, and Paraguay. The report does not regard the other two—Haiti and Myanmar—as potentially well governed, but the report calls for aid for Haiti in any case, because of its special "conflict" needs. Evidently then, the report's insistence on good government is not as rigorous as it seems at first sight.

In its zeal to produce detailed, practical aid programs, the report has gone way overboard, it seems to me, in favoring top-down development "strategies" that would place crippling administrative burdens on donors and recipient governments alike, and that are probably undesirable in any case. Sachs is convinced—and in this, he is doubtless well intentioned and sincere—that following his recommendations in full would allow the targets known as the Millennium Development Goals to be reached by 2015. Achieving these goals would mean, among other things, halving extreme poverty worldwide. Given the fact that sub-Saharan Africa is currently making no progress at all toward such a target, the report seems wildly optimistic about what might be achieved. Perhaps the aim is to inspire. But there is a danger in that—namely, a renewed cycle of exaggerated hopes and subsequent disappointment.

The best things in the report deserve nonetheless to be recognized and acted upon. Identify a group of reasonably well-governed yet very poor countries, such as those already mentioned, and gear up to deliver a lot more aid in a sustained and predictable way. Favor projects that place the fewest demands on the local political and administrative infrastructure, such as providing free anti-malarial bednets. (The report suggests a variety of straightforward actions of this sort, which it calls "quick wins.")

Very poor countries that are capable of growing faster should not be held back for lack of aid. At the moment, some are, and that is wrong. Unfortunately, however, those countries are not as numerous as the U.N. wants to believe.